View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Bradford Scholars 1 Genetic basis and timing of a major mating system shift in Capsella 2 3 Jörg A. Bachmann1,¶, Andrew Tedder1,¶, Benjamin Laenen1,¶, Marco Fracassetti1, Aurélie 4 Désamoré1, Clément Lafon-Placette2, Kim A. Steige1, Caroline Callot3, William Marande3, 5 Barbara Neuffer4, Hélène Bergès3, Claudia Köhler2, Vincent Castric5, Tanja Slotte1* 6 1Department of Ecology, Environment and Plant Sciences, Science for Life Laboratory, 7 Stockholm University, SE-106 91 Stockholm, Sweden, 2Department of Plant Biology, 8 Swedish University of Agricultural Sciences & Linnean Center for Plant Biology, SE-750 07 9 Uppsala, Sweden3Institut National de la Recherche Agronomique UPR 1258, Centre National 10 des Ressources Génomiques Végétales, Castanet-Tolosan, France, 4Department of Botany, 11 University of Osnabruck, Osnabruck, Germany, 5Univ. Lille, CNRS, UMR 8198 - Evo-Eco- 12 Paleo, F-59000 Lille, France 13 14 *Corresponding author: Email: [email protected], Phone: +46-76-1252109 15 ¶These authors contributed equally to this work. 16 17 Summary 18 • A crucial step in the transition from outcrossing to self-fertilization is the loss of 19 genetic self-incompatibility (SI). In the Brassicaceae, SI involves interaction of female 20 and male specificity components, encoded by the genes SRK and SCR at the self- 21 incompatibility locus (S-locus). Theory predicts that S-linked mutations, and 22 especially dominant mutations in SCR, are likely to contribute to loss of SI. However, 23 few studies have investigated the contribution of dominant mutations to loss of SI in 24 wild plant species. 1 25 • Here, we investigate the genetic basis of loss of SI in the self-fertilizing crucifer 26 species Capsella orientalis, by combining genetic mapping, long-read sequencing of 27 complete S-haplotypes, gene expression analyses, and controlled crosses. 28 • We show that loss of SI in C. orientalis occurred less than 2.6 Mya and maps as a 29 dominant trait to the S-locus. We identify a fixed frameshift deletion in the male 30 specificity gene SCR and confirm loss of male SI specificity. We further identify an S- 31 linked small RNA that is predicted to cause dominance of self-compatibility. 32 • Our results agree with predictions on the contribution of dominant S-linked mutations 33 to loss of SI, and thus provide new insights into the molecular basis of mating system 34 transitions. 35 36 Keywords: Capsella, dominance modifier, long-read sequencing, parallel evolution, plant 37 mating system shift, self-compatibility, S-locus, small RNA 2 38 Introduction 39 The shift from outcrossing to self-fertilization is one of the most common evolutionary 40 transitions in flowering plants (Darwin, 1876; Wright et al., 2013). This transition is favored 41 when the benefits of reproductive assurance (Darwin, 1876; Pannell & Barrett, 1998; Eckert 42 et al., 2006) and the transmission advantage of selfing (Fisher, 1941) outweigh the cost of 43 inbreeding depression (Charlesworth, 2006). 44 The transition to self-fertilization often involves breakdown of self-incompatibility 45 (SI). SI systems allow plants to recognize and reject self pollen through the action of male and 46 female specificity components and modifier loci (Takayama & Isogai, 2005). In the 47 Brassicaceae, SI is controlled by two tightly linked genes at the S-locus, the S-locus receptor 48 kinase gene SRK and SCR, which encode the female and male SI specificity determinants, 49 respectively (de Nettancourt, 2001). SRK is a transmembrane serine-threonine receptor kinase 50 located on the stigma surface (Stein et al., 1991; 1996). SCR is a small cysteine-rich protein 51 that is deposited on the pollen coat and acts as a ligand to the SRK receptor (Schopfer et al., 52 1999; Takayama et al., 2001). Direct interaction between SRK and SCR from the same S- 53 haplotype results in inhibition of pollen germination (Takasaki et al., 2000; Takayama et al., 54 2001; Ma et al., 2016) through a signaling cascade involving several proteins (Nasrallah & 55 Nasrallah 2014). This SI response prevents close inbreeding and promotes outcrossing. At the 56 S-locus, recombination is suppressed and rare allele advantage maintains alleles with different 57 specificities (Wright, 1939; Castric & Vekemans, 2004; Vekemans et al., 2014). SI 58 populations often harbor dozens of highly diverged S-haplotypes as a result of negative 59 frequency-dependent selection (Mable et al., 2003; Guo et al., 2009). In the sporophytic 60 Brassicaceae SI system, expression of a single S-specificity provides greater compatibility 61 with other individuals (Schoen & Busch, 2009). Therefore, S-haplotypes often form a 62 dominance hierarchy that determines which specificity is expressed in S-heterozygotes 3 63 (Durand et al., 2014). At the pollen level, dominance is governed by dominance modifiers in 64 the form of sRNAs expressed by dominant alleles. These sRNAs target sequence motifs 65 specific to recessive alleles of SCR, resulting in transcriptional silencing (Tarutani et al., 66 2010; Durand et al., 2014). 67 Despite the advantages of outcrossing, SI has been lost repeatedly in many different 68 lineages. There is a strong theoretical and empirical interest in the role of parallel molecular 69 changes for repeated shifts to self-compatibility (SC) (Vekemans et al., 2014; Shimizu & 70 Tsuchimatsu, 2015). While the numerous genes that act as unlinked modifiers of SI 71 potentially constitute a larger mutational target than the S-locus itself, theory predicts that 72 mutations that result in degeneration of components of the S-locus should have an advantage 73 (Porcher & Lande, 2005). Theory further predicts that the probability of spread of mutations 74 disrupting SI depends on whether they affect male or female SI function, or both functions 75 jointly (Charlesworth & Charlesworth, 1979). In particular, mutations that disrupt male 76 specificity should have an advantage over those mutations that disrupt female specificity, 77 because male specificity mutations can spread faster through both pollen and seeds 78 (Uyenoyama et al., 2001; Tsuchimatsu & Shimizu, 2013). Finally, dominant advantageous 79 mutations should have a higher fixation probability in outcrossers, as expected from 80 Haldane’s sieve (Haldane 1927). However, dominant S-alleles typically have low population 81 frequencies (Llaurens et al., 2008), resulting in a lower probability that SC mutations occur on 82 dominant than on recessive alleles. While degeneration of male specificity has contributed to 83 loss of SI in a few Brassicaceae species (Tsuchimatsu et al., 2010; 2012; Shimizu & 84 Tsuchimatsu, 2015) Chantha et al., 2013), more examples are needed. So far, few empirical 85 studies of wild species have examined the contribution of dominant S-haplotypes to the loss 86 of SI (but see Nasrallah et al. 2007). To understand the role of parallel molecular changes for 87 recurrent loss of SI, identification of causal mutations is required. This has been a challenging 4 88 task, due to the difficulty of sequencing the up to 110 kb long, highly polymorphic and 89 repetitive S-locus. However, thanks to the advent of long-read sequencing, contiguous S- 90 haplotypes can now be assembled with low error rates (Bachmann et al., 2018). 91 The crucifer genus Capsella is an emerging model for genomic studies of plant mating 92 system evolution. In Capsella, SI is the ancestral state, as there is trans-specific shared S- 93 locus polymorphism between the outcrossing SI species Capsella grandiflora and outcrossing 94 SI Arabidopsis species (Guo et al., 2009). Nevertheless, SC has evolved repeatedly in 95 Capsella, resulting in two self-compatible and highly selfing diploid species, Capsella rubella 96 and Capsella orientalis, as well as the selfing allotetraploid Capsella bursa-pastoris, which 97 formed by hybridization between C. orientalis and C. grandiflora accompanied by genome 98 duplication (Douglas et al., 2015). These species also differ greatly in their geographical 99 distributions, with C. bursa-pastoris having a nearly worldwide distribution, whereas C. 100 rubella is mainly found in Central and Southern Europe, and C. orientalis has a distribution 101 ranging from Eastern Europe to Central Asia (Hurka et al., 2012). Finally, the SI outcrosser 102 C. grandiflora mainly occurs in northwestern Greece and Albania, and in northern Italy 103 (Hurka et al., 2012). 104 In C. rubella, the transition to selfing has been intensely studied (Foxe et al., 2009; 105 Guo et al., 2009; Slotte et al., 2013; Brandvain et al., 2013) and involved the fixation of a 106 relatively dominant S-haplotype (Nasrallah et al., 2007; Guo et al., 2009; Paetsch et al. 2010) 107 most likely within the past 100-170 ky (Slotte et al., 2013; Koenig et al., 2019). Knowledge 108 on the mode, timing and demographics of the transition to selfing in C. rubella has provided 109 an evolutionary context for the study of genomic (Gos et al., 2012; Slotte et al., 2013; 110 Brandvain et al., 2013; Koenig et al., 2019), regulatory (Steige et al. 2015) and phenotypic 111 (Slotte et al., 2012; Sicard et al., 2016) consequences of selfing. In contrast, we know little 112 about the genetic basis and timing of loss of SI and transition to selfing in C. orientalis. Such 5 113 information is important for proper interpretation of genomic studies of the effects of selfing 114 and can provide insights into the role of parallel molecular changes for convergent loss of SI. 115 Here, we combined genetic mapping, long-read sequencing of S-haplotypes, 116 controlled crosses, population genomic and expression analyses to investigate the loss of SI in 117 C. orientalis, with the specific aims to: 1) test whether loss of SI maps to the S-locus, 2) 118 identify candidate causal mutations for the loss of SI, 3) investigate the role of sRNA-based 119 dominance modifiers, and 4) estimate the timing of loss of SI in C.